ROC曲线在累积比数logit判别模型效果评价中的应用

The Application of ROC Analysis in the Evaluation on Effect of Cumulative Odds Logit Discriminant Model

目的:应用ROC曲线图评价累积比数logit判别模型的效果。方法:介绍ROC分析的原理,并应用其对医学生睡眠质量的累积比数logit判别模型进行评价。结果:ROC曲线较凸,曲线下面积的可信区间不包括0.5,结论:采用ROC曲线来评价累积比数logit判别模型,综合了灵敏度和特异度两个指标,可以全面反映模型的判别效果,比较科学。

Objective： To evaluate the distinguish effect of Cumulative odds logit model with ROC curve. Methods： Introduce principle of ROC, and evaluate the model about sleep quality of medical students. Results： The ROC curve is raise, 0.5 is not included in Confidence interval of area under the curve. Conclusion： Evaluating Cumulative odds logit model with ROC curve, use two targets including sensitivity and specificity, which can reflect the model completely.